M5. Visual recognition

In this module we give to the student an overview of the latest methods based on deep learning techniques to solve visual recognition problems. The final aim is the understanding of complex scenes to build feasible systems for automatic image understanding able to answer the complex question of what objects and where are these objects in a complex scene. The students will learn a large family of successful architectures of deep convolutional networks that have been proved to solve the visual tasks of: detection, segmentation and recognition. And they are going to get the skills for designing, programming, training and evaluating complex architectures to solve vision problems.
Project title: 
Scene Understanding for automatic driving

The goal of this project is to learn the basic concepts and techniques to build deep neural networks to detect, segment and recognize specific objects, focusing on images recorded by an on-board vehicle camera for autonomous driving. 

The learning objectives are using different deep learning (DL) programming frameworks such Theano, TensorFlow and Keras and  basic DL methods such as feed forward networks (MLP) and  Convolutional  Neural Networks (CNN). It includes the understanding of standard networks for classification (AlexNet, VGG, GoogleNet, ResNet, DenseNet, SqueezeNet) detection (RCNN, Fast RCNN, Faster RCNN, YOLO) and segmentation (FCN, SegNet, UNET).  The students will learn through a project based methodology using modern collaborative tools at all stages of the project development.

The students will acquire the skills for the tasks of designing, training, tuning and evaluating neural networks to solve the problem of automatic image understanding.

Module lectures: 
Academic Year 2017-2018 under construction        
Week Date Time Lecture Lecturer University Building Room
1 Mon. Feb.25th 16:00 - 18:00 Architectures for Image Classification Adriana Romero UAB CVC Sala d'Actes
1 Mon. Feb.25th 18:00 - 19:00 Project Introduction LLuis Gómez/ Robert Benavente UAB CVC Sala d'Actes
1 Wed. Feb. 27th 16:00 - 18:00 Recurrent Neural Networks (RNN) Adriana Romero UAB CVC Sala d'Actes
               
2 Mon. Mar. 4th 16:00 - 18:00 Architectures for Object Detection Michal Drozdzal  UAB CVC Sala d'Actes
2 Mon. Mar. 4th 18:00 - 19:00 Project follow-up LLuis Gómez/ Robert Benavente UAB CVC  Sala d'Actes 
2 Wed. Mar. 6th 16:00 - 18:00 Architectures for Image Generation (GANs & VAEs)* Michal Drozdzal  UAB CVC Sala d'Actes
               
3 Mon. Mar. 11th 16:00 - 18:00 DL frameworks * Marc Masana UAB CVC Sala d'Actes
3 Mon. Mar. 11th 18:00 - 19:00 Project follow-up LLuis Gómez/ Robert Benavente UAB CVC Sala d'Actes
3 Wed. Mar. 13th 16:00 - 18:00  Semantic and Instance Segmentation Pedro Pinheiro  UAB  CVC Sala d'Actes 
               
4 Mon. Mar. 18th 16:00 -18:00 Reinforcement learning Pedro Pinheiro UAB CVC Sala d'Actes
4 Mon. Mar. 18th 18:00 -19:00 Project follow-up LLuis Gómez/ Robert Benavente UAB CVC Sala d'Actes
4 Wed. Mar. 20th 16:00 -18:00 Embedding learning: Siameses, Triplets and mining * Joan Serrat UAB CVC Sala d'Actes
               
Mon. Mar. 25th    HOMEWORK        
Mon. Mar. 27th    HOMEWORK        
               
6 Mon. Apr. 1st 16:00 -18:00 Encoder-decoder architectures * Luís Herranz  UAB  CVC  Sala d'Actes
6 Mon. Apr. 1st 18:00 -19:00 Project follow-up LLuis Gómez/ Robert Benavente UAB  CVC Sala d'Actes 
6 Wed. Apr. 3rd 16:00 -18:00

Efficient methods for Deep Learning *

Juan C. Moure  UAB CVC Sala d'Actes 
               
Mon. Apr. 8th  16:00 -18:00  Project presentations LLuis Gómez/ Robert Benavente  UAB  CVC  Sala d'Actes 
               
     

Easter Holidays

(from April 12th to 22nd)

       
               
8 Wed. Apr. 24th 18:00 -19:00 HOMEWORK        
               
9 Mon. Apr. 29th 16:00 -19:00 EXAM Joan Serrat UAB CVC Sala d'Actes
               

M5 Student Guide [here]